Sign in with
Sign up | Sign in
Your question

Computer randomly reboots

Last response: in Systems
Share
August 10, 2008 8:05:53 PM

Hello. I'm new here so I'm not sure if this is the proper place for this question. Here goes....

I have built a new system with the following specs:

Intel DX48BT2 Motherboard
Intel Core 2 Duo Quad Q9300 cpu
OCZ OCZ3p1600GK DDR3 ram
Asus EN8600GTS video card
Three 250GB barracuda hard drives
One LG DVD writer.
Two 120mm fans
One 80mm fan for the CPU duct on the case.
Windows Vista Ultimate 64 operating system.

I started with an Antec Smartpower 500w. I had random reboots, so I tried another identical one. That didn't solve it. I now have a 600w OCZ (OCZ OCZ600GXSSLI GameXStream).

I had been having random reboots (at least they seemed random). I now know I can force a reboot if I run Prime95 and it's default test ( I won't check to find out the name of it now because it WILL reboot the machine in a few minutes at most).

My core temps are around 65 centigrade normally. I have an Ultra Aluminous case with one 120mm fan in the front and one in rear. I have a CPU duct fan on the case. all three are Antec. During the tests, cpu core temps go up to 99 or so. I'm NOT an overclocker. I wouldn't know what I was doing if I tried.

What I CAN reproduce:

If I over ride my CMOS settings for ram to set the timing to 7,7,7,24 and/or set the voltage up to 1.9v, then tell it to "exit saving settings", the system simply shuts down IMMEDIATELY upon pressing "y". When I press the power button the system starts up, then shuts down for a few seconds, then restarts. It goes through the PATA detection screen then says that the POST failed and do I want to restart. It's a loop I can not exit until I put the cmos settings for ram back to "automatic", which means the timings are wrong and the voltage is low.

I've been trying everything and am about ready to pull my hair out.

I've run memtest and the ram passed the test. The machine did a reboot after a couple iterations. This runs outside WIndows, so I know it's not Vista.

Oh. Besides being on my third PSU, Intel sent me a replacement motherboard. They both behaved the same. So, It sounds like I can rule the system board and PSU out.

Thanks for whatever direction you can give me.
August 10, 2008 8:33:39 PM

Why do you want to increase memory voltage so much?

Your CPU temps are way too high or the software reports them incorrectly. What temps are reported by the motherboard's BIOS?

What CPU cooler are you using?

What happens if you temporarily remove the CPU duct? Lower or higher temps?

What's the air flow direction of the CPU duct fan?
August 10, 2008 8:52:29 PM

I got the voltage from the following link:
http://benchmarkreviews.com/index.php?option=com_conten...

I'll have to get the bios temps in a sec. I'll edit this post to add them.

The CPU cooler I'm using (I assume you mean the one mounted on the CPU), is the one that shipped with the retail box CPU. Fan and Heatsink.

The CPU duct fan is blowing OUT. If I reverse it, it'll be blowing right into the CPU fan.

Removing the side of the case so it is completely open doesn't seem to make much difference.

If I'm wrong about the RAM voltage (reading the article wrong) let me know. I'm just leaving the settings on automatic so it boots and works.

Oh, I'm using CPUFan for temp reading in Vista, as well as Everest Ultimate.
I'm looking at Everest right now:
CPU core 1 66
CPU core 2 66
CPU core 3 62
CPU core 4 63

Voltages:
CPU Core 1.10
+2.5v 1.18v
3.3v 3.25v

Memtest says
core 0: 61
core 1: 60
COre 2: 56

Doesn't list core 3. Just the three I listed.

I'll go get the CMOS temp settings and be back to append this.

Thanks.

Ok, I've checked CMOS. All I can find there relative to temp readings is the following:

Processor Therm Margin: 25c
Internal Temp 41c
Remote Temp 38c

While I was there, I wrote down the voltages:

v12 12.187
v5 4.999
v3.3 3.248
v1.5 1.184
Vccp 1.125

Am I mistaken or does V1.5 look low?

Not sure what I'm looking at, specifically.

Thanks.
Related resources
August 10, 2008 9:14:03 PM

Excuse me. In the previous, I should have said "Speedfan", not "memtest" is reporting the temps.

See? I've been messing with this so long even I'm unstable!!!

Tim
August 10, 2008 10:31:07 PM

Are you sure the cooler is installed correctly? I agree that the 1.5v seems to be low, but I can't confirm if it's OK or not.
August 10, 2008 10:42:38 PM

i would use realtemp for temp readings. Also reseat the heatsink, it doesn't sound like it was installed correctly the first time
August 10, 2008 11:58:24 PM

I'll bet you are hitting the Thermtrip# and the CPU is shutting down to save its life.

Check the HS mounting before you boot again. You may need to remove the mobo to ensure that the pins are protruding through the mobo properly.

Get a good aftermarket HS with a back plate.

Image courtesy of Grimmy.

August 11, 2008 12:29:40 AM

Well, it was doing this before I removed it and reinstalled it but I can do it again and check it out. WHen I installed it this last time, I double checked that it was seated well. I remember pushing it down again to make sure it was in all the way.

...and I replaced the thermal paste with what came with the new system board making sure I got all of it out the syringe.
August 11, 2008 12:32:58 AM

*I'll bet you are hitting the Thermtrip# and the CPU is shutting down to save its life.*

Maybe I'm splitting hairs, but it's not just shutting down, it's rebooting. The only time it shuts down (turns off completely) is when I've tried to change some ram timing values to what OCZ says they should be, then when I say "Save Bios settings and restart", it simply turns off immediately when I press "Y". Other than that, it reboots.

If that is irrelevant, fine. But I wanted to clarify that.

Thanks,

TIm
August 11, 2008 12:52:16 AM

Downloaded RealTemp and it told me that cpus 1 and 2 were in the mid sixties and 3 and 4 were in the 50's.

August 11, 2008 12:57:01 AM

timbrown527 said:
Well, it was doing this before I removed it and reinstalled it but I can do it again and check it out. WHen I installed it this last time, I double checked that it was seated well. I remember pushing it down again to make sure it was in all the way.

...and I replaced the thermal paste with what came with the new system board making sure I got all of it out the syringe.




To much paste will make temps to high.

It takes a dot the size of a BB or a grain of rice. Using the whole syringe would be very bad.
August 11, 2008 1:00:50 AM

You wrote "...and I replaced the thermal paste with what came with the new system board making sure I got all of it out the syringe. " You want to apply a thin layer on the CPU, not the whole syringe contents. If it's the "Dow Corning TC-1996 compound Intel P/N D54816-001", the syringe contains enough compound to install several CPUs.
August 11, 2008 1:04:27 AM

I would hate to clean up this mess.
August 11, 2008 1:05:54 AM

Ok, I have opened the box, loosened the "screws" on the cpu fan assembly and pushed on all four posts. I pushed as hard as I could without fearing damage to the board. One post budged a bit but that's it. I'm running the temp program again and it's telling me the same thing.

As far as tearing the system down and removing the system board to check it, I'll have to wait till tomorrow to do that. No big deal, just remove the system tray.

August 11, 2008 1:15:44 AM

*Get a good aftermarket HS with a back plate.*

Can you recommend one?

August 11, 2008 1:17:31 AM

*It takes a dot the size of a BB or a grain of rice. Using the whole syringe would be very bad. *

That makes more sense to me than thinking it's not seated all the way. I think the local radio shack has this, I'll re-do it.

Thanks.
August 11, 2008 8:42:10 AM

timbrown527 said:
*I'll bet you are hitting the Thermtrip# and the CPU is shutting down to save its life.*

Maybe I'm splitting hairs, but it's not just shutting down, it's rebooting. The only time it shuts down (turns off completely) is when I've tried to change some ram timing values to what OCZ says they should be, then when I say "Save Bios settings and restart", it simply turns off immediately when I press "Y". Other than that, it reboots.

If that is irrelevant, fine. But I wanted to clarify that.
That is relevant. I don't believe that the machine will reboot if the thermtrip# is reached. After you work out the temp issue, assuming it's a temp issue and not a DTS issue, and I think it is, we will need to tackle the crashing issue.

timbrown527 said:
Ok, I have opened the box, loosened the "screws" on the cpu fan assembly and pushed on all four posts. I pushed as hard as I could without fearing damage to the board. One post budged a bit but that's it. I'm running the temp program again and it's telling me the same thing.

As far as tearing the system down and removing the system board to check it, I'll have to wait till tomorrow to do that. No big deal, just remove the system tray.
The stock push pin heatsink is notoriously hard to install and the pins sometimes work themselves back out. If you remove it to reapply the paste, be sure to clean both mating surfaces thoroughly with rubbing alcohol and a coffee filter.
timbrown527 said:
*Get a good aftermarket HS with a back plate.*

Can you recommend one?
There are many good ones. A lot people also like some of the aftermarket HSs with the push pins, I am not one of them. Those HSs also suffer from the poor push pin design of the stock cooler, and many have complained about this and converted their aftermarket push pin HS to a backplate one.

I have a Thermalright Ultra120 Extreme with a Scythe S-FLEX SFF21F 120mm Fan. It is a little older and more expensive, but it's still one of the best. I leave the fan at max (1600 rpm) and it's not too loud, although I'm not really the best judge of noise.

Here is a relatively recent comparison list of HSs, I'm sure there are more recent ones. Noctua NH-U12P CPU Heatsink Review

The list is pretty complete. Rather than recommend one, I'll let you decide. Your case should be deep enough for any of the HSs, though with the tray you may want to double check clearance. Stick to a heatsink with a backplate. Heatsinks with a proprietary fan may present a problem if the fan dies.

timbrown527 said:
*It takes a dot the size of a BB or a grain of rice. Using the whole syringe would be very bad. *

That makes more sense to me than thinking it's not seated all the way. I think the local radio shack has this, I'll re-do it.
Although too much grease is bad for temps, and messy, I don't believe it is responsible for the ultra high temps you have. This assumes that your DTSs aren't completely whacked. Most of the excess paste should get squeezed out as the HS is applied, depending on its consistency, and hopefully you wouldn't have gotten any voids.

I like IC Diamond 7 Carat Thermal Compound, although no one else at THF seems to be very excited about it. I apply a very thin layer with a single sided razor blade because it is a little sticky. It can be heated to thin it out a little and applied in the standard fashion. It can also be purchased from the Heatsink Factory. If you decide to use it, or are bored, this is a good thread to read, especially posts by JoeC as it appears he works for Innovation Cooling.
August 11, 2008 10:29:07 AM

Thanks.

I may have been a bit premature, but I called TigerDirect and ordered a Thermaltake. Here's the link.
http://www.tigerdirect.com/applications/searchtools/ite...

IT should be here tomorrow. I ordered thermal paste with it.

I'm replacing the heatsink/fan that came with the chip mainly because while I removed it, one of the posts actually broke off. Thanks for the input on how to clean the thermal paste off, I didn't know how to do that properly.

In any event, that system is down until tomorrow. I'll use a coffee filter and alcohol to clean the top of the processor off. then put a small amount of paste on it.

When I removed the fan, the paste had NOT all been spread out in a thin film. It had been on one side, but not on the side that had the broken fan post. So, it would appear that at least some of the heat problem was due to a contact issue.

NOTE: When I first assembled this machine it had been randomly rebooting once every week or two. When I remomved the heatsink, the paste WAS evenly distributed, so I assume that contact was good on the first install. And of course, the past on the fan had been put there by Intel so the amount should have been right.

I'm guessing I cracked the push pin when I installed the cpu fan; I've been told I don't know my own strength.

Finally, what do you mean by a "DTS problem"? Never heard of this before.

Thanks to you, and all of you, for your help on this. I have built many systems over the years but this one has been the one that "refused to be built". The last one i built was about three or four years ago, a plain old 2.5Ghz P4. Things change.

Again, I appreciate your help very much.
August 11, 2008 1:34:23 PM

It is too bad that you jumped the gun and ordered that HS, especially because it has the same mounting pins that the stock one has, and is therefore subject to the same problems. Additionally there are other HS available for within $10-$15 that I suspect would crush it in cooling performance. The reason that I say suspect is that I can't even find it included in any reviews, that should be an indication. What's done is done.

The DTS (Digital Thermal Sensors) are the on chip thermal sensors that are used in the newer Intel and AMD CPUs. Many of the Intel 45nm CPUs do a poor job of reporting idle temps. Intel says this is because they were designed only for protection so only the high temps are of concern to them. A little bit of a cop-out here? Also Intel doesn't release the Tjmax numbers for the various CPUs so it's somewhat of a crap shoot to determine them. They are needed because the core temp is calculated by subtracting from the Tjmax. Also there can be some problem with them sticking, but I haven't heard a problem as severe as yours.

As I indicated earlier, I suspected that the HS wasn't getting the proper contact. It's not the first time that I have had someone insist that the HS was properly mounted only to find out that it wasn't. I strongly suggest that you remove the tray and ensure that the pins are firmly seated and even give a "little" tug on the HS to ensure that it won't work loose later.

Core 2 Quad and Duo Temperature Guide
Real Temp - Documentation
Real Temp - New temp program for Intel Core processors - XtremeSystems Forums






August 11, 2008 1:47:14 PM

Zorg said:
It is too bad that you jumped the gun and ordered that HS, especially because it has the same mounting pins that the stock one has, and is therefore subject to the same problems. Additionally there are other HS available for within $10-$15 that I suspect would crush it in cooling performance. The reason that I say suspect is that I can't even find it included in any reviews, that should be an indication. What's done is done.

The DTS (Digital Thermal Sensors) are the on chip thermal sensors that are used in the newer Intel and AMD CPUs. Many of the Intel 45nm CPUs do a poor job of reporting idle temps. Intel says this is because they were designed only for protection so only the high temps are of concern to them. A little bit of a cop-out here? Also Intel doesn't release the Tjmax numbers for the various CPUs so it's somewhat of a crap shoot to determine them. They are needed because the core temp is calculated by subtracting from the Tjmax. Also there can be some problem with them sticking, but I haven't heard a problem as severe as yours.

As I indicated earlier, I suspected that the HS wasn't getting the proper contact. It's not the first time that I have had someone insist that the HS was properly mounted only to find out that it wasn't. I strongly suggest that you remove the tray and ensure that the pins are firmly seated and even give a "little" tug on the HS to ensure that it won't work loose later.

Core 2 Quad and Duo Temperature Guide
Real Temp - Documentation
Real Temp - New temp program for Intel Core processors - XtremeSystems Forums


Yeah, I know, I jumped a bit quick. I ordered it last night and called them this morning about it. I was going to get a (better) one but they said it was already on the truck. I would think I can trust that because they could have sold a more expensive one if the current one wasn't already on its way.

I noticed some of the reviews about it were negative (pins) but I did see some good ones and figured I'd not worry about it. I can always change later, or do an RMA if this one isn't sufficient. I would think it's somewhat better than the one from Intel.

I'll check the ones out you mentioned. I can always Ebay the one I'm getting, so it's no great loss, and at least it'll let me run till I get a better one.


August 11, 2008 1:50:56 PM

*As I indicated earlier, I suspected that the HS wasn't getting the proper contact. It's not the first time that I have had someone insist that the HS was properly mounted only to find out that it wasn't. I strongly suggest that you remove the tray and ensure that the pins are firmly seated and even give a "little" tug on the HS to ensure that it won't work loose later. *

I'm sure you are right because of the way the past looked when I took the fan off. it was not "smooshed" on one side of the CPU (the side where the pin was cracked).

When I installed it, I did so with it on the tray, out of the case but I didn't remove the system board to look at the back. I DID try to look at it through the crack between the tray and the board, but didn't see and in my rush didn't dismount the board.

August 11, 2008 2:08:48 PM

It's definitely better than the stock one. Just make sure it's mounted securely.

There are a few schools of thought for applying the TIM (Thermal Interface Material).

1. Coat the whole surface with a very very thin coating.
2. Use a grain of rice sized drop in the middle.
3. Lay it down in a line. How to apply Arctic Silver 5 - Intel Quad Core CPU W/Heatspreader

I use #1, but I think most like #3.
August 11, 2008 2:17:02 PM

I thought there was a hole over the CPU socket. Since that's not the case, then you really should remove the mobo from the tray, many are successful without doing that. You will need to support the mobo so that the pins have clearance, and so that the mobo doesn't crack if you need excessive force. A small roll of masking/electrical tape etc. around the hole will do the trick. Be sure it is soft enough and not resting on anything that might be damaged e.g., protruding legs of capacitors etc.
August 11, 2008 3:23:35 PM

wow!!!! what a battle i hope that the OP and others have learned a lesson,,always mount those foolish and stupid push pin junky things outside of your case,and check before installation that it is firmly in place..
y'know if i were working at inhell and someone had came to me with that push pin idea,,well pilgrim,, there aint enough paint ball ammo on the entire planet,talk about stupidity,,, whew!! .:>)
August 11, 2008 3:42:36 PM

zorg:

WIll do. And I'll let you know how it goes when the job is done sometime tomorrow afternoon.

You've been a great help. Thanks.
August 11, 2008 4:05:50 PM

Damn right man, the guys in this forum are geniuses - I was also a complete amateur and managed to work out how to setup a class newbuild in a day (although i haven't actually started yet...) thanks to their help.
I hope the solution works man, but if it doesn't, don't worry because you're getting better tech support than any crappy helpline/p.c. shop could ever hope to give......
I hope it works!
August 11, 2008 4:17:50 PM

I appreciate the confidence, but let's not get carried away here. ;) 

I just posted this in another thread. This is the hot HS right now, and it's cheap. I didn't mention it earlier because it comes with, God help me, push pins stock, but there is a back plate that fits it.

Zorg said:
XIGMATEK HDT-S1283 120mm Rifle CPU Cooler ONLY WITH THIS XIGMATEK ACK-I7751 Retention Bracket
It's a knock off of the Ultra 120 and it actually performs better for cheap as long as you have the bracket. Also you need to put the thermal paste on the heat-pipes and base, because it will fill the cracks and not spread properly if you put a dab in the center of the CPU.
August 12, 2008 7:48:39 PM

Well, it would appear that I won't be able to report back with results for a few more days. TigerDirect decided (for some reason) to ship the fan UPS BASIC instead of ground. Ground usually is next day. BASIC is three to four (according to one agent. . .another one said up to seven).

Guess from now on I'll have to say "SHIP IT GROUND!"
August 15, 2008 9:35:02 PM

Ok, it would appear I'm back up and running. I had ordered the heat sink/fan unit and the pins DID prove to be a waste of time. I returned it, but they sent another one instead of one with screws. Today I FINALLY got the one I ordered, screws, backplate and all. It's a low end thermaltake but at least my temps are down. As of right now:

CPU Core 1: 52c
CPU Core 2: 52c
CPU Core 3: 48c
CPU Core 4: 48c

Those are all lower than before by 20c or so.

August 16, 2008 12:14:27 AM

I'm running Real Temp right now and the TJMax numbers are as follows:

49, 49, 54, 54

The temps in the tray are 46, 46, 42, 42.

WHen I had RealTemp test the sensors, it reported that perhaps one or more of the sensors are stuck.


Thanks,

TIm
August 19, 2008 12:02:41 AM

Well, the rebooting apparently isn't a temperature issue. I'm sitting here with the cores running in the mid forties and it just rebooted.

What do we look at next?

August 19, 2008 3:07:36 AM

Anything obvious in Event Viewer? It should provide some indication as to the reason why the system rebooted.
Anonymous
August 19, 2008 3:32:46 AM

once again... zorg pwnage...

zorg whats better about the diamond paste compared to the AS 5 or MX-2 thermal paste?
August 20, 2008 1:34:29 PM

GhislainG said:
Anything obvious in Event Viewer? It should provide some indication as to the reason why the system rebooted.



No. Event Viewer has never reported anything past "Unexpected shutdown".

I've contacted OCZ about the ram. They are working with me on it.
September 10, 2008 5:20:22 AM

I ran into this problem on two DX48BT2 motherboards. One running a x64 OS and the other a x86 OS. Same problem on both machines. In the x86 I'm running 4G of Corsair and on the x64 machine I'm running 8G of Patriot. On both machines I was getting the random re-boots and on the x64 machine it wouldn't even start. The problem is the lack of power to the RAM. I've upped the memory voltage to 1.7 and upped the CPU voltage to 1.175. On the x64 machine I had to go with 1.86 on the 8 Gigs. Everything works great now! Also make sure that your timing is at 9-9-9-24 and at 1333/333mhz (if your ram supports it). I have successfully installed over 10 more of these motherboards within our network and all work great with the voltage increase. If this doesn't work let me know. I've got extensive time invested in these motherboards.


!